Why Use MCP?

@playbykey/theory-mcp exists to give developers and musicians an intuitive engine to build with, and to give AI assistants the same thing: a way to compute music theory relationships instead of predicting them.

Why it matters

AI models can still hallucinate a music theory fact outright, or misread which key or notation convention a question assumes: a relative minor off by a third, a chord with the wrong enharmonic spelling, a mismatched roman numeral, transposition off by a semitone crossing an octave. This does the most damage on fundamentals. If you’re learning and an AI confidently gets something basic wrong, you may not catch it immediately.

The same risk applies to anything you build. This engine lets you power your AI with fundamentals, instead of just being AI-powered.

A foundation to build on

This engine computes fixed relationships: chords, scales, progressions, transposition, key relationships, MIDI and frequency conversion. You can:

• Ask an AI assistant a music theory question and get a consistent answer, grounded in the actual theory
• Build a new music tool, app, or learning platform with this as the foundation
• Add reliable music theory grounding to something you’ve already built

PlayByKey itself is built on this foundation.

Example prompts and use cases

Ask an AI assistant directly:

• “What are the diatonic chords in D major?”
• “Show me the inversions of a Cmaj7 chord”
• “What’s the I-V-vi-IV progression in G major?”
• “What’s the relative minor of F major?”
• “Transpose this progression from C major to A major”
• “What is the MIDI note number for A4?”

Or build with it:

• A practice app calls it to generate quiz answers grounded in the actual theory
• An AI tutor calls it before answering, so what it explains is correct, not just plausible-sounding
• A production tool or plugin calls it for key and chord data it can depend on

Setup

See playbykey/mcp for installation across Claude Desktop, Cursor, and Claude Code, plus the full tool reference.